采用多层感知器神经网络构建亚急性期缺血性脑卒中患者短期预后的预测模型

doi:10.3969/j.issn.1006-9771.2022.03.009

《中国康复理论与实践》 ›› 2022, Vol. 28 ›› Issue (3): 335-339.doi: 10.3969/j.issn.1006-9771.2022.03.009

采用多层感知器神经网络构建亚急性期缺血性脑卒中患者短期预后的预测模型

赖海芳¹,顾琳²,纵亚¹,牛传欣¹,谢青¹()

1. 上海交通大学医学院附属瑞金医院,上海市 200025
2. 上海市瑞金康复医院,上海市200023

收稿日期:2021-10-11 修回日期:2021-11-05 出版日期:2022-03-25 发布日期:2022-03-31
通讯作者: 谢青 E-mail:ruijin_xq@163.com
作者简介:赖海芳(1991-),女,汉族,福建福鼎市人,硕士研究生,主治医师,主要研究方向：神经系统疾病的康复治疗。
基金资助:
上海市临床重点专科项目(shslczdzk02701);上海交通大学医学院附属瑞金医院(北部院区)研究基金项目No(2020ZY10)

Prediction of short-term outcome after subacute ischemic stroke using multiple layer perceptron neural network

LAI Haifang¹,GU Lin²,ZONG Ya¹,NIU Chuanxin¹,XIE Qing¹()

1. Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai 200025, China
2. Shanghai Ruijin Rehabilitation Hospital, Shanghai 200023, China

Received:2021-10-11 Revised:2021-11-05 Published:2022-03-25 Online:2022-03-31
Contact: XIE Qing E-mail:ruijin_xq@163.com
Supported by:
Shanghai Municipal Key Clinical Specialty No(shslczdzk02701);Ruijin Hospital, Shanghai Jiaotong University School of Medicine (Northern Campus) Research Fund No(2020ZY10)

摘要/Abstract

摘要： 目的采用多层感知器(MLP)神经网络构建亚急性期缺血性脑卒中(CIS)短期预后的预测模型。方法 2019年1月至2021年9月,上海市瑞金康复医院康复科再住院缺血性脑卒中患者60例,采集首次住院时(病程< 30 d)的临床相关信息,根据首次入院3个月后改良Rankin量表评分,判断患者短期预后。采用单因素分析筛选与短期预后相关的危险因素,分别采用常规多因素Logistic回归和MLP建立预测模型,计算两种模型的预测准确率,采用接受者操作特征(ROC)曲线评估预测效应。结果多因素Logistic回归模型预测准确率73.3%,ROC曲线下面积0.851;MLP模型预测准确率88.9%,ROC曲线下面积0.930。结论采用MLP模型能更好预测亚急性期缺血性脑卒中的短期预后。

关键词: 缺血性脑卒中, 亚急性期, 多因素Logistic回归, 多层感知器, 神经网络, 预测, 短期结局

Abstract: Objective To establish a predictive model using multiple layer perceptron (MLP) for short-term outcome after subacute ischemic stroke.Methods From January, 2019 to September, 2021, 60 readmission-inpatients in Department of Rehabilitation, Ruijin Hospital, Shanghai Jiaotong University School of Medicine were collected the clinical features of first admission (less than 30 days after attack), and the outcomes were assessed with modified Rankin Scale (MRS) three months after the first admission. The risk factors were screened with single factor analysis, and the short-term outcome predictive models were established with multi-factor Logistic regression and MLP. The predictive accuracy of both models was calculated, and the predictive effects were compared with Receiver Operating Characteristic (ROC) curve.Results For multi-factor Logistic regression, the predictive accuracy was 73.3%, and the area under ROC curve was 0.851. For MLP, the predictive accuracy was 88.9%, and the area under the ROC curve was 0.930.Conclusion The prediction of short-term outcome after subacute ischemic stroke can be done with MLP model.

Key words: ischemic stroke, subacute, multi-factor Logistic regression, multiple layer perceptron, neural network, prediction, short-term outcome

赖海芳,顾琳,纵亚,牛传欣,谢青. 采用多层感知器神经网络构建亚急性期缺血性脑卒中患者短期预后的预测模型[J]. 《中国康复理论与实践》, 2022, 28(3): 335-339.

LAI Haifang,GU Lin,ZONG Ya,NIU Chuanxin,XIE Qing. Prediction of short-term outcome after subacute ischemic stroke using multiple layer perceptron neural network[J]. 《Chinese Journal of Rehabilitation Theory and Practice》, 2022, 28(3): 335-339.

图/表 7

表1

表2

表3

表4

图1

表5

图2

参考文献 25

[1]	WANG W Z, JIANG B, SUN H X, et al. Prevalence, incidence, and mortality of stroke in China: results from a nationwide population-based survey of 480687 adults[J]. Circulation, 2017,135(8):759-771.
[2]	D'AMORE C, PACIARONI M, SILVESTRELLI G, et al. Severity of acute intracerebral haemorrhage, elderly age and atrial fibrillation: independent predictors of poor outcome at three months[J]. Eur J Intern Med, 2013,24(4):310-313.
[3]	SEO W K, SEOK H Y, KIM J H, et al. C-reactive protein is a predictor of early neurologic deterioration in acute ischemic stroke[J]. J Stroke Cerebrovasc Dis, 2012,21(3):181-186.
[4]	KUNICKI T J, WILLIAMS S A, NUGENT D J, et al. Mean platelet volume and integrin alleles correlate with levels of inte-grins alpha (IIb) beta (3) and alpha (2) beta (1) in acute coronary syndrome patients and normal subjects[J]. Arterioscler Thromb Vasc Biol, 2012,32(1):147-152.
[5]	宋祖玲, 刁莎, 严兰平, 等. 联合决策树及Logistic回归建立乳腺癌相对风险预测模型[J]. 现代预防医学, 2019,46(7):1156-1175.
	SONG Z L, DIAO S, YAN L P, et al. Prediction model of relative risk for breast cancer based on decision tree and Logistic regression[J]. Moder Prev Med, 2019,46(7):1156-1175.
[6]	TU J V. Advantages and disadvantages of using artificial neural networks versus logistic regression for predicting medical outcomes[J]. J Clin Epidemiol, 1996,49(11):1225-1231.
[7]	RENGANATHAN V. Overview of artificial neural network models in the biomedical domain[J]. Bratisl Lek Listy, 2019,120(7):536-540.
[8]	郑广勇, 蔡剑. 多层感知器神经网络在手足口病预测模型中的应用[J]. 中国公共卫生管理, 2020,36(1):67-73.
	ZHENG G Y, CAI J. Application of multi-layer perceptron neural network in the prediction model of hand-foot-mouth disease[J]. Chin J PHM, 2020,36(1):67-73.
[9]	LAU K W K, MAK M K Y. Speed-dependent treadmill training is effective to improve gait and balance performance in patients with subacute stroke[J]. J Rehabil Med, 2011,43(8):709-713.
[10]	王冬燕, 吕航. 神经网络方法在医学统计预测中的应用[J]. 南京中医药大学学报(社会科学版), 2017,18(1):47-52.
	WANG D Y, LÜ H. Application of neural network method to medical statistics forecast[J]. J Nanjing Univ Tradi Chin Med (Soc Sci), 2017,18(1):47-52.
[11]	KEIDAR D, YARON D, GOLDSTEIN E, et al. COVID-19 classification of X-ray images using deep neural networks[J]. Eur Radiol, 2021,31(12):9654-9663.
[12]	CHETHAM S M, BARKERT M, STAFFORD W. Neural networks in cardiac electrophysiological signal classification[J]. Australas Phys Eng Sci Med, 2002,25(3):124-131.
[13]	LING T S, WU L L, FU Y W, et al. A deep learning-based system for identifying differentiation status and delineating the margins of early gastric cancer in magnifying narrow-band imaging endoscopy[J]. Endoscopy, 2021,53(5):469-477.
[14]	KNOFLACH M, MATOSEVIC B, RÜCKER M, et al. Functional recovery after ischemic stroke: a matter of age: data from the Austrian Stroke Unit Registry[J]. Neurology, 2012,78(4):279-285.
[15]	KUGLER C, ALTENHONER T, LOCHNER P, et al. Hessian Stroke Data Bank Study Group ASH. Does age influence early recovery from ischemic stroke? A study from the Hessian Stroke Data Bank[J]. J Neurol, 2003,250(6):676-681.
[16]	CHALOS V, VAN DER ENDE N A M, LINGSMA H F, et al. National Institutes of Health Stroke Scale: an alternative primary outcome measure for trials of acute treatment for ischemic stroke[J]. Stroke, 2020,51(1):282-290.
[17]	AHMED R, ZUBERI B F, AFSAR S. Stroke scale score and early prediction of outcome after stroke[J]. J Coll Physicians Surg Pak, 2004,14(5):267-269.
[18]	RAWTHER T, TABET F. Biology, pathophysiology and current therapies that affect lipoprotein(a) levels[J]. J Mol Cell Cardiol, 2019,131:1-11.
[19]	GENCER B, KRONENBERG F, STROES E S, et al. Lipoprotein(a): the revenant[J]. Eur Heart J, 2017,38(20):1553-1560.
[20]	ERQOU S, KAPTOGE S, PERRY P L, et al. Lipoprotein(a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality[J]. JAMA, 2009,302:412-423.
[21]	FORBES C A, QUEK R G, DESHPANDE S, et al. The relationship between Lp (a) and CVD outcomes: a systematic review[J]. Lipids Health Dis, 2016,15:95.
[22]	ZHANG W, ZHANGX A. Prognostic value of serum lipoprotein (a) levels in patients with acute ischemic stroke[J]. Neuroreport, 2014,25(4):262-266.
[23]	FARAJI F, GHASAMI K, TALAIE-ZANJANI A, et al. Prognostic factors in acute stroke, regarding to stroke severity by Canadian Neurological Stroke Scale: a hospital-based study[J]. Asian J Neurosurg, 2013,8(2):78-82.
[24]	WORTHMANN H, TRYC A B, GOLDBECKER A, et al. The temporal profile of inflammatory markers and mediators in blood after acute ischemic stroke differs depending on stroke outcome[J]. Cerebrovas Dis, 2010,30(1):85-92.
[25]	WANG A X, TIAN X, ZUO Y T, et al. High lactate dehydrogenase was associated withadverse outcomes in patients with acute ischemic stroke or transient ischemic attack[J]. Ann Palliat Med, 2021,10(10):10185-10195.

特征变量	数值
总数/n	60
性别(男/女)/n	36/24
年龄/岁	67.3±10.3
高血压病/n	48
糖尿病/n	25
吸烟史/n	13
饮酒史/n	10
CRP/mg·L^-1	11.5±14.3
白蛋白/g·L^-1	34.1±3.4
空腹血糖/mmol·L^-1	6.1±1.8
脂蛋白a/mg·L^-1	24.8±32.2
D-D聚体/mg·L^-1	0.5±0.8
B型钠尿肽/pg·mL^-1	257.7±437.7
入院NHISS评分	7.5±4.6

因变量	n	预后良好数	χ²值	P值
性别			1.111	0.292
男	36	16
女	24	14
高血压史			1.667	0.197
有	48	22
无	12	8
糖尿病史			0.069	0.793
有	25	12
无	35	18
吸烟史			0.884	0.347
有	13	5
无	47	25
饮酒史			0.218	0.640
有	10	4
无	50	26

特征变量	预后良好 (n = 30)	预后不良 (n = 30)	F值	P值
年龄/岁	63.53±10.94	71.13±9.37	8.35	0.005
CRP/mg·L^-1	7.43±2.22	15.57±19.35	5.23	0.026
白蛋白/g·L^-1	34.98±3.17	33.25±3.40	4.16	0.046
空腹血糖/mmol·L^-1	6.33±1.91	5.90±1.67	0.86	0.358
脂蛋白a/mg·L^-1	15.12±15.74	34.61±40.82	5.97	0.018
入院NHISS评分	5.57±2.76	9.43±5.30	12.55	0.001
D-D聚体/mg·L^-1	0.31±0.45	0.64±0.97	2.81	0.099
B型钠尿肽/pg·mL^-1	266.32±215.04	145.17±155.35	2.23	0.123

自变量	B	S	P值	OR	95%CI
年龄	0.086	0.040	0.031	1.090	1.008~1.178
CRP	0.112	0.082	0.172	1.118	0.953~1.313
白蛋白	-0.030	0.112	0.787	0.970	0.779~1.208
脂蛋白a	0.028	0.017	0.096	1.029	0.995~1.064
入院NHISS评分	0.205	0.091	0.024	1.228	1.027~1.467
常量	-8.989	3.044	0.003	0.000

模型		良好/n	不良/n	准确率/%
Logistic回归
良好		23	7	76.7
不良		9	21	70.0
合计				73.3
MLP
训练	良好	23	0	100.0
	不良	2	17	89.5
	合计			95.2
检验	良好	7	0	100.0
	不良	2	9	81.8
	合计			88.9

采用多层感知器神经网络构建亚急性期缺血性脑卒中患者短期预后的预测模型

Prediction of short-term outcome after subacute ischemic stroke using multiple layer perceptron neural network

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 25

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	张瑞明, 王凯, 尹锐, 王海亮, 莫言. 失重条件下舱外手套外骨骼助力效果的评估方法[J]. 《中国康复理论与实践》, 2023, 29(7): 856-861.
[2]	邓婷, 陈敬绵, 刘小蒙, 姚晓华, 刘芦姗, 何威, 张通, 芦海涛. 轻中度急性缺血性脑卒中并发卒中相关性肺炎的危险因素分析[J]. 《中国康复理论与实践》, 2023, 29(6): 708-713.
[3]	张春龙, 刘福亮, 商娜, 李芳, 刘慧珍. 血清脂联素、超敏C-反应蛋白与急性缺血性脑卒中短期预后的相关性[J]. 《中国康复理论与实践》, 2023, 29(10): 1221-1226.
[4]	张博超,杨朝,郭立泉,陈静,熊大曦. 基于机器学习的慢性阻塞性肺疾病急性加重预测模型的研究[J]. 《中国康复理论与实践》, 2022, 28(6): 678-683.
[5]	王博,袁永学,张庆苏. 非经口进食吞咽障碍脑卒中患者预后的相关因素及预测模型[J]. 《中国康复理论与实践》, 2022, 28(4): 453-460.
[6]	俞快,张利,叶祥明. 慢性意识障碍患者结局预测模型的系统综述[J]. 《中国康复理论与实践》, 2022, 28(2): 190-198.
[7]	朱慧,夏有兵,巩尊科,王世雁,马柯,闫金秋. 高频重复经颅磁刺激对缺血性脑卒中后中枢性面瘫的效果[J]. 《中国康复理论与实践》, 2022, 28(2): 199-203.
[8]	张倩,卞敏洁,何琴,黄东锋. 血管性认知障碍早期预测机器学习模型的构建[J]. 《中国康复理论与实践》, 2021, 27(9): 1072-1077.
[9]	王娜,李培兰,刘芦姗,王丰容. 不同评分系统预测急性缺血性脑卒中患者的短期和长期死亡风险比较[J]. 《中国康复理论与实践》, 2021, 27(3): 256-260.
[10]	王娜娜, 杜小正, 何文洁, 张馨予, 郑欣, 李梦莘. 针刺治疗缺血性脑卒中的Notch信号通路研究进展[J]. 《中国康复理论与实践》, 2021, 27(1): 67-70.
[11]	陈小平,王明洋,马登磊,龚诗立,张丽,李雅莉,李林,胡朝英,张兰. 痘苗病毒致炎兔皮提取物对脑缺血再灌注脑损伤大鼠凋亡和神经炎症的影响[J]. 《中国康复理论与实践》, 2020, 26(9): 1038-1044.
[12]	刘建华,魏清川,胡秀茹,叶赛青,闫志宇,高强. 弥散张量成像各向异性分数与缺血性脑卒中患者上肢运动功能的相关性[J]. 《中国康复理论与实践》, 2020, 26(7): 749-752.
[13]	魏海萍,郭佳,王欢,葛朝明. 迷走神经刺激对脑缺血再灌注大鼠磷酸腺苷活化蛋白激酶-沉默信息调节因子2相关酶1自噬通路的影响[J]. 《中国康复理论与实践》, 2020, 26(7): 775-779.
[14]	刘慧珍,郭树彬,商娜,李芳,刘芦姗,李培兰,陈敬绵,王丰容,李俊玉. 血清25羟基-维生素D对急性缺血性脑卒中急诊患者结局的预测价值[J]. 《中国康复理论与实践》, 2020, 26(7): 830-835.
[15]	张燕,王铭玥,王婕,姜恺宁,张筠晗. 基于Xception-LSTM的下肢运动能力评价方法[J]. 《中国康复理论与实践》, 2020, 26(6): 643-647.